THE EARTHQUAKE ADVANCE THEORY 175 observed in Yakutat Bay were quite different in degree and in character from those of other regions. Hypothesis of Climatic Variation. The consideration of the problem naturally began •with the commonly-accepted theory for fluctuations in glacier fronts—response to variations in snowfall in the glacier reservoirs. No one questions the rationality of this theory when applied to such slow advances and recessions as occur, for instance, in the Alps, where for a series of years ice fronts advance or recede slightly; but it is a question worthy of careful consideration whether this cause is also competent to explain the spectacular changes observed in the Yakutat Bay region. In considering this question it is to be borne in mind that we have to account for the sudden advance of nine or ten glaciers in a limited region, and that these glaciers vary in size and in the conditions of their reservoirs. Under the theory of climatic variations it would be necessary to assume that after a period of a half century, or thereabouts, of snowfall not quite sufficient to hold glacier fronts in their position, thus giving rise to stagnation in their outer portions, and accompanying recession, the snowfall abruptly increased and enough fell to produce these startling phenomena. Then came a cessation of the excessive snowfall as abrupt as the beginning. In other words, in a limited area there must have been a great increase in snowfall for a few years and then a return to the normal. It must be insisted, too, that the phenomena of advance show clearly that the transformation could not be accounted for by moderate increase in snowfall. Only a very notable addition to the reservoirs could bring about so spasmodic and so great a forward movement. In the hope of getting some light on the question of variation in snowfall in this region we have examined the precipitation data for stations on the Alaskan coast within a distance of 250 miles to the southeast and an equal distance to the northwest.1 These records, though fragmentary, prove, what is otherwise well known, that there are great variations in amount of precipitation from place to place along this coast, and since the nearest of these stations is 150 miles distant from Yakutat Bay, it is evident that no safe deduction can be drawn from the recorded precipitation in an interpretation of variation in snowfall on the slopes of the St. Elias Range. We do not know whether the precipitation on the mountain slopes is greater or less than at Sitka or Nuchek, nor do we know how it varies with elevation or exposure, nor whether an increase in precipitation at Sitka was associated with an increase on the slopes of the St. Elias mountains. After a careful consideration of the question we have become convinced that no definite or satisfactory discussion of the subject is possible on the basis of existing records of precipitation. On one point, however, there is noteworthy testimony from the records, namely that, for some reason, there is, at least locally, a very decided variation in precipitation. If the records are accurate, there has been at Sitka, 250 miles southeast of Yakutat, a variation in rainfall of from 59 inches in 1872 to 140 inches in 1886. From 1848 to 1876 the average precipitation was 79 inches, and in no year did it rise above 94 inches, while in only two years did it fall below 65 inches. Up to this time the precipitation was fairly regular, but a precipitation of 77 inches in 1876 was followed by one i Based upon records made by the Russians and published in, House Ex. Doc. 177, 40th. Congress, 2nd Session, Washington, 1868, pp. 884-fi, 387; upon U. S. Weather Bureau records published by Cleveland Abbe, Jr., n The Geography and Geology of Alaska, Professional Paper 45, U. S. Geol. Survey, 1906, pp. 188-200; and upon later manuscript records furnished us by the U. S. Weather Bureau.